Thread Rating:
  • 1 Vote(s) - 5 Average
  • 1
  • 2
  • 3
  • 4
  • 5
miniEngine version 2 - idea thread
#41
Dear Hans,

thanks for you ideas. Let me go through them...

(07-31-2013, 12:42 PM)hansai Wrote: I have another suggestion or wish.
For demonstration purpose I would like to have a demo mode.

When the keyframes are successfully implemented, I guess it is pretty easy to define some kind of demo. This would basically just load a predefines set of keyframes and run them.

(07-31-2013, 12:42 PM)hansai Wrote: A second variant would be a given move and return to Home with maximum speed and it starts again until the users stops the process.
Therefore you would need a max. speed menu point. This has to be adjusted just once by the user in order of the max. steps/sec. possible by his motor or setup.

The max-speed menu point is currently available in the original miniE and will definitely be in the new version too. I can imagine that a specific keyframe exists which just moves the motor back to home and restarts the program.

(07-31-2013, 12:42 PM)hansai Wrote: A external start button would be a great additional feature. It could be connected with a contrast detection unit. This would give the opportunity to start the slider when a action starts.

The new function - external trigger - could be used for that. In the end it is just a question of configuration what function the trigger will have. This is definitely on the list!

Thanks again for your ideas!

Cheers,
A.
Reply
#42
Is there a mailing list for version 2 when it comes out?
I will like to be in the list to order it as soon as you make it available.
thanks.
I know your system is more sophisticate but if you do not mind
could you take a look at this guy code ?
http://www.schweinert.com/blog/files/e0777d515ab541e921d4bbffd17e3726-16.html
Is for video but I manage to get it working and I like it.
thanks again
Reply
#43
(08-09-2013, 12:37 AM)efelectric Wrote: Is there a mailing list for version 2 when it comes out?

Hello efelectric,

there is no mailing list but you could follow my twitter account or subscribe to the RSS feed of my blog.

I post every major and some minor updates of the project progress there.

Cheers,
Airic
Reply
#44
I like your design there are a couple of stepper drivers that maybe good for version 3 :-)

I like that way the L6470 chip takes care of a lot of the processing freeing up the microprocessor to do other calculations.
http://altelectronics.co.uk/shop/arduino/arduino-stepper-motor-shield/prod_84.html

I just came across this shield and I know from their TNCN-343 board they produce a good product.
http://www.motioncontrol-community.org/?page_id=353

Waiting for your boards to become available for V2
Reply
#45
Hey Greg,

Thanks for the links! The Arduino DUE actually has plenty of processing power to do the job for two motors simultaneously without any problems. As mentioned in my blog, I got the core motor routines done and they allow to control the precise definition of when the motor should be where. Through the function-based algorithms even the current speed of the motor is known (cm/sec) so measurements can be done which motor/driver combination develop the best performance (speed vs torque).

The internal calculations also allow much deeper tuning of the actual moves because every aspect of the curve can be tweaked (if wanted). It also allows the combination of the miniEngine 2 with any stepper driver that suits the requirements for the job.

Cheers,
Airic
Reply
#46
Airic,
I'm a happy user of miniE 1.3 and I see that 2.0 is going to be a great tool.
I find that you adding SD slot to be able to save some settings.
I have maybe a crazy idea how to use this slot.
Calculation of Bézier curves is quite processor demanding. Do you think that it will be possible to calculate on desk top computer all curves (you thinking about 2 but 5 (3D + focus + zoom) would cover all needs for time-lapse control) by dedicated software and save it as incremental steps vs previous clock point matrix on SD card)
miniE 2.0 could just read from the card next moves without calculation of each step.
I'm not sure if such approach is possible or it is workable and user friendly and of course controlling 5 motors is far beyond your project but I hope you can consider this option
regards
Reply
#47
(10-23-2013, 01:25 PM)robot66 Wrote: I'm not sure if such approach is possible or it is workable and user friendly and of course controlling 5 motors is far beyond your project but I hope you can consider this option
regards

Hello robot66,

the Bezier curve algorithms are actually done and working fine. The general approch is to pre-segmentate the Bezier curves into linear segments. As linear segments are linear functions, the problem with the CPU-heavy calculations was shifted to a RAM-heavy problem. One segmentation (200 segments per curve) takes something between 15ms and 150ms (based on the algorithm used). So when starting the run, the preparation of the moves should not take more than 1-2 seconds. As I will integrate a start delay anyway (for letting the systme settle after powering on all motors), this comes in perfectly.

The existing RAM of the DUE is capable of holding at least 25 of these curves which should be well enought for setting up a reasonable complex move with two motors. Every daisy-chained unit brings 2 more motors and also its memory for another 25 curves. Please check my blog (http://minie.airiclenz.com) for more information about this topic.

Cheers,
A.
Reply
#48
Dear Airic,
a new suggestion.
When it comes to a remote head with two or three axis, it could be a good idea to have as a first step in configuring this system with a given lens or dimension, kind of a software stop for movements.
I mean, set the home position - pan forward to a save position where you dont touch anything - push a button and this position is secured as most forward possible movement - do the same with backwards and tilt.
There will not be a head which could turn every camera with every long lens 360° in both directions. Use kind of end buttons like on the slider is useless on such a system.
These settings should overrule other settings allways.
Reply
#49
Hello hansai,

nice idea - I added it to the psec-list.

Ceers,
Airic
Reply
#50
Hello Airic,

Just started using your board, think it is great, a new suggestion or maybe I missed it.

Support for 2 servos for a pan/tilt head, servo moves saved on SD card:
PT785-S pan and tilt system
http://www.servocity.com/html/pt785-s_pan___tilt_system.html

I modified this project to work using 2 servos, stepper motor with all moves saved on a SD card. On the SD card I input the time, date, slider stepper move distance, camera and servo pan and tilt movement for each shot. This setup allows for some very long time-lapse periods.

Long Term Time-Lapse Photography with Arduino and CHDK
http://interface.khm.de/index.php/lab/experiments/time-lapse-photography-with-arduino-and-chdk/

Thanks for looking
Gary
Reply


Forum Jump:


Users browsing this thread: 3 Guest(s)